Comparison of Devices That Measure Sickle Red Cell Deformability

Abstract

Background With the emergence of novel pharmacologic and gene-based therapies, identifying the rheological and biophysical RBC abnormalities of sickle cell disease (SCD) not captured by clinical laboratory techniques is crucial. Ektacytometry (LORRCA), the current standard to assess RBC function, does not reflect mechanical stress that RBCs experience in capillary microvasculature in the body. The device is costly, requires a nitrogen gas tank, and its use is difficult to translate outside of large-scale laboratory settings. In comparison, the Microfluidic Impedance Red Cell Assay (MIRCA) is a low-cost, portable device that mimics capillary microvasculature with micropillar arrays spaced 3-12 µm apart. Thus, MIRCA measures mechanical deformability, which is more physiologic, mimicking RBCs squeezing through microvasculature. An impedance analyzer calculates the MIRCA Occlusion Index (OI), representing the % occlusion of the chip. Here we compare the MIRCA to the LORRCA, assessing the correlations of OI and Elongation Index maximum (EI) to conventional laboratory tests and SCD related complications. Methods Peripheral blood from 53 adult (n = 28) and pediatric (n = 25) individuals (HbSS = 35, HbSB0 = 6, HbSC = 5, HbSB0 = 1, and HbAA controls = 6) were obtained under an Emory University IRB approved protocol. Patients transfused < 90 days prior were excluded. Hospitalizations / emergency department (ED) visits for pain within 12 months of sample collection were determined by chart review. Samples were collected in EDTA, stored at 4°C up to 48 hours, centrifuged at 500g, washed, and resuspended to a 20% hematocrit to run on MIRCA, an oxygen gradient ektacytometer (LORRCA), and an ADVIA hematology analyzer. Polydimethylsiloxane (PDMS) fabricated MIRCA chips were bonded to a standard glass slide with pairs of gold-sputtered electrodes adjacent to each array. Chips were prepared by perfusing ethanol, 1X phosphate-buffered saline (PBS), and 3% bovine-serum albumin (BSA) in 1X PBS via syringe pump and were incubated overnight at 4°C prior to use. Baseline impedance was taken for a 2-minute perfusion of 1X PBS, then OIs were calculated from impedance values 10 minutes after sample introduction normalized to baseline. The data was analyzed using the Mann-Whitney U test, Spearman correlation, and linear regression. OriginPro (Northampton, MA, USA) and Stata 18 (College Station, TX, USA) were used for the analyses and a p < 0.05 was considered significant. Results The correlation coefficients for OI and EI with laboratory tests (Table 1) and to each other (Figure 1) were comparable (p < 0.05). HbSS/SB0 had higher OIs than HbSC/SB+ (median = 13.7% vs 6.8%, p < 0.01) and HbAA (median = 4.6%, p < 0.01). Patients with ≥ 1 vaso-occlusive events (VOE) in the past year had higher OIs and lower EIs compared to those without a VOE (median = 15.3% vs 9.6%, p <0.01; median = 0.4 vs 0.5, p = 0.02, respectively). Patients that received acute medical care (admission or ED visit) had higher OIs and lower EIs compared to those that did not (median = 15.7% vs 9.8%, p < 0.01; median = 0.4 vs 0.5, p = 0.02, respectively). In multiple linear regression, HbSS/SB0 (p < 0.01), adults (p = 0.01), and DRBC (p < 0.01) were associated with OI while HbSS/SB0 (p = 0.01), DRBC (p < 0.01), hemoglobin % (p = 0.04), absolute neutrophil count (p = 0.02), and VOE+ (p = 0.02) were associated with EI. The models explained 55.2% and 66.8% variability of OI and EI, respectively. Lower Akaike's information criterion (AIC, 186 vs -80) and Bayesian information criterion (BIC, 192 vs -71) of the EI model suggest a higher correlation to conventional laboratory values and clinical features compared to OI; OI may possibly be a unique biomarker that both laboratory tests and LORRCA fail to fully characterize. Conclusion MIRCA OI and LORRCA EI normoxic deformability are correlated to each other, and both show statistically comparable correlations to many parameters associated with SCD severity. Both OI and EI are significantly associated with SCD outcomes like VOE and acute care. Due to lower device size, cost, and relative ease-of-use, MIRCA may be more convenient for routine clinical use and for use in low resource settings where LORRCA costs and requirements hamper implementation. As the addition of chemical hypoxia to MIRCA is now underway, future work will examine OI under hypoxia and its associations with SCD-relevant outcomes like pain events, stroke and acute chest syndrome.